This makes the cross section
to be tested 2" X 3". Some of the blocks are cut radially and
some tangentially, as indicated in the figure.

[Illustration: FIG. 45.--Design of cleavage test specimen.]

The free ends of the hooks are fitted into the notch in the end
of the specimen. The movable head of the machine is then made to
descend at the rate of 0.25 inch per minute, pulling apart the
hooks and splitting the block. The maximum load only is taken
and the result expressed in pounds per square inch of width. A
piece one-half inch thick is split off parallel to the failure
and used for moisture determination.

TENSION TEST PARALLEL TO THE GRAIN

Since the tensile strength of wood parallel to the grain is
greater than the compressive strength, and exceedingly greater
than the shearing strength, it is very difficult to make
satisfactory tension tests, as the head and shoulders of the
test specimen (which is subjected to both compression and shear)
must be stronger than the portion subjected to a pure tensile
stress.

Various designs of test specimens have been made. The one first
employed by the Division of Forestry[61] was prepared as
follows: Sticks were cut measuring 1.5" X 2.5" X 16". The
thickness at the centre was then reduced to three-eighths of an
inch by cutting out circular segments with a band saw. This left
a breaking section of 2.5" X 0.375". Care was taken to cut the
specimen as nearly parallel to the grain as possible, so that
its failure would occur in a condition of pure tension. The
specimen was then placed between the plane wedge-shaped steel
grips of the cage and the movable head of the static machine and
pulled in two. Only the maximum load was recorded. (See Fig. 46,
No. 1.)

[Illustration: FIG. 46.--Designs of tension test specimens used
in United States.]

[Footnote 61: Bul. No. 8: Timber physics, Part II., 1893, p. 7.]

The difficulty of making such tests compared with the minor
importance of the results is so great that they are at present
omitted by the U.S. Forest Service. A form of specimen is
suggested, however, and is as follows: "A rod of wood about one
inch in diameter is bored by a hollow drill from the stick to be
tested. The ends of this rod are inserted and glued in
corresponding holes in permanent hardwood wedges. The specimen
is then submitted to the ordinary tension test. The broken ends
are punched from the wedges."[62] (See Fig. 46, No. 2.)

[Footnote 62: Cir.